Prediction AI compute, hardware and data

Hassabis said he thinks Google DeepMind will have the most compute by far of any research lab in 2024.

Unresolvable The public record could not settle it.

Prediction: Statements about events outside the person's control, checked once they fell due: did they come true?

The second judge would not resolve this claim; the verdict shown is the first judge's.

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

Judged over calendar 2024. True: credible third-party estimates (for example, from Epoch AI, SemiAnalysis or similar analysts) or official disclosures show that the AI compute available to Google DeepMind (including the Google TPU/GPU capacity allocated to it) in 2024 was the largest of any AI research lab, and 'by far': a clear and large lead over the next lab (for example, roughly 1.5x or more of OpenAI, Meta AI, xAI, Anthropic, Microsoft and others). Mostly true: it had the most compute with a clear lead, but the lead was somewhat smaller than 'by far' implies. Mostly false: it was roughly tied with another lab, or it led only on a narrow measure. False: credible estimates show at least one other lab (for example, Meta or OpenAI with Microsoft) had clearly more compute in 2024. If estimates conflict so much that no ranking can be made, the claim is unresolvable rather than false.

Set before the research began.

Evidence (4)

E1 Five hyperscalers now own over two-thirds of global AI compute Epoch AI · 14 April 2026

“2024 Q4 2024-12-31 Google 1.4e+6 20.9 2024 Q4 2024-12-31 Microsoft 1.2e+6 18.6 2024 Q4 2024-12-31 Meta 7.7e+5 11.7”
E2 Leading AI companies have hundreds of thousands of cutting-edge AI chips Epoch AI · 9 October 2024

“Note that our “Google” estimate includes all of Alphabet.”
E3 Estimates of GPU or equivalent resources of large AI players for 2024/5 — LessWrong LessWrong (post by CharlesD) · 28 November 2024

“I get to numbers in the range of 1m-1.5m H100 equivalents as of year end 2024.”
E4 Leading AI companies have hundreds of thousands of cutting-edge AI chips Epoch AI · 9 October 2024

“This is the equivalent of ~650,000 H100s, assuming that the TPU fleet overall has the same energy efficiency as the TPUv4.”

Reasoning

The only ranking in the evidence is an Alphabet-wide figure with no stated unit, which shows Google about 1.17x ahead of Microsoft at the end of 2024 (E1, E2). It does not show the compute available to Google DeepMind itself, and it does not settle whether that lead counts as clear or as roughly tied under the criteria.

Cite this claim

Demis Hassabis, “Hassabis said he thinks Google DeepMind will have the most compute by far of any research lab in 2024.” (said 28 February 2024). Verdict: unresolvable. Reliability API, claim c_97156f89f7f0, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_97156f89f7f0

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

More prediction claims by Demis Hassabis

Resolved true Prediction 23 July 2025

“I think fusion and solar are the two that I would bet on.”

Asked whether it would be nuclear fusion, Demis Hassabis said he thinks fusion and solar are the two energy sources he would bet on.
Resolved true Prediction 28 February 2024

“I think the next versions of this over the next year, 18 months, we’ll maybe have some contextual understanding of the environment around you through a camera or a phone or some glasses.”

Hassabis said he thought the next versions of these systems, over the next year to 18 months, might have some contextual understanding of the user's environment through a camera, phone or glasses.

All of them

Verdicts are made automatically; how verdicts are made.